The goal of the proposed research is to use information from event-related brain potential (ERP) recordings to analyze the special processes that are invoked when a strong context or expectation is violated. The utility of the ERP approach in this endeavor is attested to by the numerous studies which have described the sensitivity of the amplitude of the late positive "P300" complex to fluctuations in a person's expectations. Recent data from our laboratory, however, have indicated that not all unexpected events are followed by a P300. In particular, we have found that the infrequent violation of a semantic expectancy for a word at the end of visually sented sentence is associated with a late, monophasic negative wave (N400) which is distinct from the P300. We proposed that the N400 might reflect a delay in the ongoing processing triggered by the semantic incongruity and the subsequent backtracking to the point of interruption ("second look") and the "reprocessing" which attempts to make sense of the situation. The proposed studies are aimed at testing this hypothesis and at defining the conditions wherein some violations of expectancy are associated with the P300 complex and others the N400. The following questions will be addressed specifically: 1. How responsive is the N400 to variations in the parameters of the reading situation, such as (a) the probability of the semantic anomaly, (b) the amount or strength of context, (c) the serial position of the anomaly in the sentence, (d) the rate of word presentation and (e) the interval between the context and the anomalous word? 2. Is the N400 specifically associated with violations of semantic expectancy or will it occur following violations of well-known grammatical and/or lexical rules as well? If N400 waves are found in response to other linguistic anomalies, will their properties be specific to the type of violations? 3. Are semantic expectancies modality specific as some authors have contended, or modality independent? That is, will an N400 occur in response to a semantic discrepancy at the end of an aurally presented sentence in the same way as in the silent reading task? Moreover, will the "N400 effect" occur even when the semantic incongruity and the preceding context are presented in different modalities? 4. How dependent is the elicitation of the N400 upon the reading task in which it was originally observed? Would similar ERPs be obtained in other situations requiring a semantic analysis, e.g., the sentence-picture verification task of Clark & Chase (1972)? 5. Is the N400 uniquely related to the violation of linguistic rules, or will it appear in response to unexpected notes in well-known musical sequences? 6. And finally, are the N400 and P300 lateralized to the informed cerebral hemisphere in "split-brain" patients? What can we learn about the cognitive processes underlying the P300/N400 dichotomy by mapping and comparing their distributions in the normal and split-brain individuals?